Blind manual resend
A new event ID hides the connection to the original failure and weakens duplicate protection.
CRM webhook failure recovery
Recover a failed CRM webhook without losing or duplicating the lead.
Use an explicit failed-delivery replay that preserves the original event identity and verifies payload and destination fingerprints.
A successful automation run is not proof of delivery.
Manual recovery often means copying lead fields into a new request after the incident. If the original request actually arrived late, that new request can create a duplicate. If sensitive payloads are retained indefinitely for convenience, the recovery queue becomes a privacy liability.
Failed receipt -> original payload resubmission -> fingerprint check -> CRM webhook retryChanged recovery payload
Editing the lead or destination during replay makes it impossible to prove the retry matches the failed event.
Sensitive recovery queues
Complete lead payload storage expands the data that must be secured and governed.
Implementation
Put LeadProof in the delivery path.
Keep the tools that create and process the lead. Standardize only the fragile handoff between them.
Open sandbox instructions →- Keep the original source payload in the authorized source system, not in LeadProof.
- Use the failed event's original Idempotency-Key and destination.
- Add X-LeadProof-Replay: true and resubmit the identical body.
- Confirm the recovered status and original receipt identity in the customer dashboard.
POST
https://leadproof.jessesay.chatgpt.site/api/v1/leadsAuthorization: Bearer lp_live_your_key
Idempotency-Key: stable-source-lead-id
Content-Type: application/json
{
"destination": "https://your-crm.example/webhook",
"name": "Alex Morgan",
"email": "alex@example.com",
"source": "crm recovery"
}Questions
What teams ask before adding LeadProof.
Why does replay require the original payload?
LeadProof intentionally does not retain the complete production body, so the authorized source must resubmit it.
What happens if the replay body changed?
LeadProof rejects the request with an idempotency conflict before delivery when the stored fingerprint does not match.
Does a replay consume another plan event?
A verified replay of the same stored event does not increment the new-event usage count.
See the gaps in your real workflow.
Get an explainable risk score, prioritized fixes, and the right LeadProof plan.